Occasional papers - The Museum, Texas Tech University

PUBLICATIONS OF THE MHSETM
TEXAS TECH UNIVERSITY

Three publications of The Museum of Texas Tech l^ni-
\ersity are issued under tlie auspices of the Dean of the
Craduate School and Director of Academic Publications, and
in cooperation with the International Center for Arid and
Semi-Arid Land Studies. Short researdi studies are published
as Occasional Papers whereas longer contributions appear as
Special Publications. Pajxas of practical apjrlication to col¬
lection management and museum operations are issued in
the MusTOlogy series. All are numlx'red separately and pub¬
lished on an irregular basis.

The preferred abbreviation for citing The Museum’s Occa¬
sional Papers is Occas. Papers Mus., Texas Tech Univ.

Institutional sidrscripiions are available through Texas
Tech Press, Texas Tech Ihiiversity, Lubbock, Texas 79-409.
Institutional libraries interested in exchanging publications
should address the Exchange Librarian at Texas Tech Ihii¬
versity. Individuals can purchase separate numbers of the
Occasional Papers for $1.00 each from Tt*xas Tech Press.
Remittance in IhS. currency check, money order, or bank
draft must lx- enclosed with request (add $1.00 per title or
200 pages of publications requested for foreign postage; resi¬
dents of the state of Texas must pay a 5 per cent sales tax on
the total purchase price). Copies of the “Revised checklist of
North American mammals north of Mexico, 1979" (Jones et
al., 1979, Occas. Papers Mus., Texas Tech Univ., 62:1-17) are
available at 60 cents each in orders of 10 or more.

ISSN 0149-175X
Texas Tech Press
Lidibock, Texas 79409

7£rx

OCCASIONAL PAPERS
THE MUSEUM

MUS. COMP. ZOOL..

LIBRARY

JAN 1 41Gb 1

HARVARD

TEXAS TECH UNIVERSITY'""""'^

NUMBER 73

9 JANUARY 1981

SYSTEMATICS OF THE SOUTHERN RACES
OF TWO SPECIES OF KANGAROO RATS
(DIPODOMYS COMPACTUS AND D. ORDII)

George D. Baumgardner and David J. Schmidly

Populations of kangaroo rats (Rodcntia: Heteromyidae: Dipod-
omys) occur in the southern portion of Texas, south of a line
from Del Rio to San Antonio to Aransas Pass, and on the adja¬
cent barrier islands of Texas and Tamaulipas, Mexico. These
populations, as arranged taxonomically by Hall and Kelson
(1959), comprise four subspecies of the wide ranging species
Dipodomys ordii. However, recent genetic and phenetic evidence
(Johnson and Selander, 1971; Brownlee, 1973; Best and Schnell,
1974; and Stock, 1974) suggests that kangaroo rats from the barrier
islands of Texas and Mexico (referred to Dipodomys compactus)
are specifically distinct from populations on the South Texas
mainland (referred to Dipodomys ordii). Schmidly and Hendricks
(1976) found compactus to differ markedly from adjacent races of
ordii in certain cranial features. More im[)ortantly, they discovered
that both species occur on the Sf)uth lexas mainland, apparently
without interbreeding. These authors found the two species
within 15 miles of one another in Willacy Gounty, and all tren¬
chant morphological and chromosomal differences were
maintained.

Subsequently, we have located three additional sites of s\mpati\
in Zapata and Jim Hogg counties. The purpose of this paper is to
assess morphological differentiation between compactus and ordii
with particular reference to these sympatric sites, \anation within
and among populations of compatus as well as southern popula-

2

OCCASIONAL PAPERS MUSEUM TEXAS TECH UNIVERSITY

lions (Texas and Mexico) of ordii is reviewed also in order to
allocate South Texas specimens to recognized subspecific taxa.
Schmidly and Hendricks (1976) recognized four subspecies of D.
compactus, namely, D. c. compactus, from Padre Island; D. c. lar-
gus from Mustang Island; D. c. parvabullatus from the barrier
islands of Tamaulipas; and D. c. sennetti from the South Texas
mainland. These authors referred South Texas samples of ordii to
D. o. durranti, although this assignment was tentative because
they had not examined specimens of all adjacent subspecies.

Methods and Materials

We examined 629 specimens, recording external measurements
from specimen labels (TL, total length; TAL, tail length; HFL,
hind foot length) as well as the following cranial measurements
(abbreviations before each character are used hereafter): GSL,
greatest skull length; MW, maxillary width; LIW, least interor¬
bital width; GSW, greatest skull width; RW, rostral width; NT,
nasal length; LMTR, length maxillary toothrow; LSW, least
supraoccipital width; LMB, length mastoid bulla; WMB, width
mastoid bulla; SD, skull depth; WSP, width supraoccipital at
suture; and WI, width interparietal at suture. Cranial measure¬
ments were taken according to Desha (1967), with a few excep¬
tions, and skull depth was recorded as described by Hooper
(1952:10). Width of the supraoccipital (WSP) was taken across this
bone from one junction of the supraoccipito-parietal suture and
the mastoid bulla to the other. Width of the interparietal (WI)
was measured from one junction of this bone and the
supraoccipito-parietal suture to the other. Specimens were aged
according to the method outlined by Desha (1967), and only
adults were used in statistical analyses.

Specimens were examined from 205 localities. These were plot¬
ted on a map and subsequently combined into 63 samples (Figs.
1, 2), each having enough specimens to yield meaningful statistics
and small enough in aerial extent to include potentially inter¬
breeding populations in a relatively homogenous environment.
Both compactus and ordii occur sympatrically at three sites (41-42,
44-45, and 48-49), and at these each species was considered a
separate sample. The locality for each specimen included in a
sample group is given in specimens analyzed.

Univariate analyses of the data were performed using two sub¬
routines (Procedure Means and Procedure Anova) of the Statistical
Analysis System (SAS). Procedure Means generates standard statis-

BAl’\lC;ARnNKR AM) SCHMIDl.V — kAN(;AR()() RAIS

.1

Fig. 1. — Geographic localities of Dtpodomys rompaclus (sipiares) and /). ordu
(dots) from Fexas and Mexico. Ciron|)ed samples used in the statistical analyses are
outlined and nnmlx'ied.

tics (mean, range, standard deviation, standard error of the mean,
variance, and coefficient of variation — C\'). When comparing two
or more groups. Procedure Anova tests for significant differences
(P<0.05) among the means of the groups by employing a single
classification analysis of variance.

4

OCCASIONAL PAPERS MUSEUM TEXAS TECH UNIVERSITY

Fig. 2. — Geographic localities of Dipodomys compactus (squares) and D. ordii
(dots) in South Texas. Triangles represent localities where both species were col¬
lected. The northern boundary of the Rio Grande Plain is depicted as a broad line
along the northern and eastern edge of the area.

Several multivariate statistical techniques were employed. The
Numerical Taxonomy Programs (NT-SYS) of Rohlf and Kish-
paugh (1972) were used to cluster samples according to phenetic
affinity (using average taxonomic distance as a measure of sim¬
ilarity and the UPGMA cluster option). Matrices were generated
from both the standardized character means and the unstandard¬
ized canonical variable means for those vectors that accounted for
a significant amount of variation (characteristic roots > 1).

BAUM(,ARDNF,R AND SCiHMIDI.V — RANC.AROO RAIS

Higher cophenetic correlation values were obtained from distance
phenograms generated with the unstandardized canonical variable
means, and only these phenograms are illustrated and discussed.
Canonical means were derived frc^mi a multivariate analysis of var¬
iance (MANOX’A) program in SAS.

A MANOV'A-canc:)nical analysis also was used to assess the
degree of divergence among samples. Discriminant function anal¬
ysis was used tc^ determine the extent to which reference samples
of two different comparisons (South Texas ordii and compactus]
mainland cornpactus and island compactus) could be distin¬
guished from one another. Detailed explanations of these statisti¬
cal techniques are given in Schmidly and Hendricks (1976), Yates
and Schmidly (1977), Honeycutt and Schmidly (1979), and Wil¬
kins and Schmidly (1979).

Nongeographic Variation

Sexual variation. — No consistent sexual dimorphism was found
in any of the characters analyzed at 13 separate localities. Width
of the mastoid bulla showed the greatest difference between sexes
but it was significantly different (ANOVA: P<.05) at only three of
the 13 localities. Males of compactus w^ere slightly larger than
females in external measurements, but females were larger in most
cranial features. Males of ordii were only slightly larger than
females. Because differences between males and females were not
statistically significant in most characters, sexes were combined
for subsequent statistical analyses.

I'hese results do not agree with those of Desha (1967), Schmidly
(1971), and Kennedy and Schnell (1978) who reported extensive
sexual dimorphism in samples of D. ordii. However, those three
studies were limited either to a few pop^idations (Desha, 1967;
Schmidly, 1971) or covered geographic areas not considered in this
study (Kennedy and Schnell, 1978). Schmidly and Hendricks
(1976) included both males and females in an analysis of geogra¬
phic \ariation and noted only limited sexual dimorphism in cra¬
nial and external measurements of kangaroo rats from South
Texas and Mexico.

Individual variation. — Coefficients of variation (CA') for exter¬
nal measurements of compactus (sample .51) ranged from 1.82
(HFL, male) to 5.49 (TL, male); for cranial measurements, from
1.34 (GSL, female) to 15.32 (WI, male). C\'s for external mea¬
surements of D. ordii (sample 35) ranged from 3.38 (HkL, male)
to 6.55 (TL, male); for cranial measurements from 1.40 (NL,

6

OCCASIONAL PAPERS MUSEUM TEXAS TECH UNIVERSE! Y

female) to 22.3 (WI, male). Males of both compactus (mean
CV=5.28) and ordii (mean CV=6.44) are slightly more variable
than females (mean CVs=3.59 and 5.05, respectively).

The characters LSW, WSP, and WI exhibited high CVs. This
coidd result from a lack of refinement in taking these measure¬
ments or from the influence of bullar inflation, which reduces
these cranial elements from two directions (Lidicker, 1960).
Although the reliability of characters involving the supraoccipital
and interparietal bones has been questioned by several authors
(Lidicker, 1960; Schmidly, 1971), some acceptable measurement of
these features is useful in determining the extent of bullar infla¬
tion. CVs of these measurements are near the upper limits of
those considered acceptable for taxonomic studies by Long (1968).
For compactus, LSW exhibits the lowest mean CV (9.94); for
ordii, WSP (mean CV=11.34) has the lowest value.

Pelage variation. — Island populations of compactus exhibit
intrapopulational variation in color. Two distinct color phases
have been recorded in samples of this species, namely Light
Ochraceous-Buff (red) and Cartridge Buff (gray) (Setzer, 1949).
The frequency of these two color phases in four island popula¬
tions is as follows (per cent incidence of gray phase followed by
per cent incidence of red); Mustang Island, 17.2, 82.8; N Padre
Island, 65.3, 34.7; S Padre Island, 93.3, 6.7; and Tamaulipas, 5.6,
94.4.

Geographic Variation
Univariate Analysis

To investigate the distinctness of compactus with respect to
ordii a west to east transect was constructed for five characters
(GSL, GSW, LMB, WMB, WSP) among 14 samples from main¬
land South Texas (Fig. 3). Characters that reflect the width of the
supraoccipital (WSP) and the size and inflation of the auditory
bulla (GSW, LMB, WMB) separated the samples into two groups,
a compactus group (samples 49, 41, 44, 43, 51, 53, 46, 54) and an
ordii group (50, 48, 42, 45, 47, 52). Differences were reinforced by
comparisons made between samples where the two species occur
sympatrically— the characters GSW, LMB, and WMB differed sig¬
nificantly between groups at these localities. Only GSL showed
general overlap among all samples.

Patterns of univariate variation among samples of D. compac¬
tus were examined along a transect proceeding from Mustang
Island, Texas, south to the barrier islands of Tamaulipas, Mexico,

BAl'MCiARDNER AND SC.HMIDI A — kAN(,AR()() RA IS

7

LOC. N

GSL

GSW

LMt

WMI

WSf

Fig. 3. — I’nivariate variation, expressed by Dite-Leraas diagrams of selected
characters, among samples of Dipodomys cornpactus (stippled closed rectangles)
and D. ordii (dark closed rectangles) along a transect in South Texas. Sample
designation (Loc) and size (\) appear along the left margin (if N varies, the differ¬
ent N appears to the right of the diagram). See Fig. 1 for location of samj)les. Fhe
horizontal line represents the range; vertical line, the mean; open rectangle, one
standard deviation; and closed rectangle, two standard errors of the mean.

and thence northwest across mainland South Texas to Jim Hogg
County. The Dice-Teraas diagrams do not reveal a iinift:)rm pat¬
tern of geographic variation among island samples (Fig. 4). Spec¬
imens from south Padre Island (59) are the largest while those
from Famaulipas (22) are the smallest. To the northwest, along
the mainland, size increases slightly; mainland samples are
slightly larger than island forms in three characters (GSW, TMB,
WMB), although these differences are not statistically significant.

Patterns of univariate variation among samples of D. ordu were
examined for five characters (CiSL, MW, TMB, WMB, SD) along a
transect extending from Oklahoma southwest to Chihuahua, and
thence southeast to San Luis Potosi (Fig. 5). Proceeding along
this transect, a reduction in size is evident in all charac ters. A dis¬
tinct break sep)arates samples from Oklahoma and northern I exas
(1, 2, 5), which are significantly largc'i in all characters except
WMB and TMB from southern sam})les of ordu. Size gradually
decreased beginning with samples from I rans-Pecos, I exas, on
into southern Chihuahua. Ciontinuing southeastward into north¬
ern San Luis Potosi, size remained fairly constant ((»SL, MW, SD)
or increased slightly (LMB, W'MB). Proceeding southwaid fiom
here, a general decrease in size was evident. second transcTt
encompassing samples from South I exas and mc:)\ing westwaid

8

OCCASIONAL PAPERS MUSEUM TEXAS TECH UNIVERSITY

LOC

N

33

28

36

3

38

10

59

24

22

8

58

4

54

7

46

10

51

25

41

23

GSL

MW

FWn

GSW

CMC

33 35 37 39 17 19 21 23

21

23

NL

LMB

WMB

33

28

36

3

38

10

59

24

22

8

58

4

54

7

46

10

51

25

41

23

Pdc.

4.—

n

OB

12

14

16 12

14

16

ae

8

10

11

Leraas diagrams of selected characters, along a transect through its range. See F'igs.
1 and 2 for location of samples and Fig. 3 for an explanation of the diagrams.

into central Mexico (35, 42, 47, 15, 18, 21) also was analyzed, but
the Dice-Leraas diagrams showed little significant variation and,
for that reason, were not illustrated. Samples from South Texas
are similar to those in southwestern Coahuila.

Multivariate Analyses

Cluster analysis.— k distance phenogram using all samples was
generated, which showed two major groupings (cophenetic corre¬
lation coefficient, 0.926). Cluster A corresponded to samples of
compactus and cluster B to samples of ordii. Within the ordii
cluster, samples 1-5 (from Oklahoma and northern Texas) formed

BAl’M(;ARnNKR AND SCHMIDI.V — KAN(;AR()() RA IS

9

toe

H

OSl

MW

LM«

WMt

2

10

1

IS

S

25

7

32

6

40

10

15

•

25

14

12

17

13

It

24

19

12

20

7

21

30 ‘

23

» i

24

9 I

25

20 1

27

20 1

im ]

c~«t.
L -mb-]

35 37 39 41 17 19 21 23 15 17

10 11

so

13 15

Pig. 5. — Geographic variation in Dtpodomys ordii, expressed by Dice-Leraas
diagrams of selected characters, along a transect from Oklahoma to central Mexico.
See Pig. 1 for location of samples and Fig. 3 for an explanation of the diagrams.

a separate and distinct subgroup. Samples of compactus (41, 44,
49) and ordii (42, 45, 48) from the three sites of sympatry in Jim
Hogg and Zapata counties fell within their proper species group¬
ing but were clustered together at fourth and third order levels,
respectively.

7'o elucidate patterns of variation within the compactus and
ordii clusters, each was subjected to separate cluster analysis
(excluding samples 1-5 of ordii). The compactus samples (Fig. 6A)
are arranged in two clusters. Cluster A includes those samples
from the barrier islands plus mainland sample 58; cluster B
represents all samples from the South Texas mainland except 58.

The phenogram for ordii (Fig. 6B) also separates into two
groups (C and D), with the exception of sample 55 (Willacy
County, Texas), which segregates by itself. Group (C) consists of
specimens from VV'est Texas (sample 6) and adjacent Chihuahua,
Mexico (9, 11). Ciroup (D) contains the remaining samples of
ordii and can be further divided into subgroups I and II. Sub¬
group I includes samples from western Fexas (7). southern Fexas
(32), northern Mexico (12), and southern Mexico (26). Subgroup
II consists of the remaining samples from the Big Bend Basin and
Rio Grande Plain of Fexas and the Mexican Plateau. I his sub¬
group separates into northern and southern divisions, a and b,
respectively. The northern division represents samples from the
Mesa del Norte of Mexico (10, 13-20), the Big Bend Basin (8), and
all South Texas ordii, whereas southern di\ision includes samples
from the Mesa Central of Mexico (21, 23-25, and 27).

10

OCCASIONAL PAPERS MUSEUM TEXAS TECH UNIVERSITY

24 .20 .17 .13 .10 .06 .03 -.004

I - 1 - 1 - 1 - 1 - 1 - 1 - 1

I - 1 - 1 - 1 - 1 - 1 - 1 - 1

.13 .11 .09 .07 .05 .03 .01 -.002

Fig. 6. — Distance phenograms of the cluster analyses for samples of Dipodomys
compactus (A) and D. ordii (B) analyzed separately. The cophenetic correlation
coefficient for the compactus cluster is 0.889; for ordii, 0.8.57.

BArM(;ARl)\F.R AND SCHMIDI.V — KAN(,AR()() RAIS

II

M A N () I’ A -carionica I analysis. — Four cliffereiit criteria
( Hotelling- Lawley’s Trace, Pilla’s Trace, Wilks’ Caiterion, and
Roy’s Maxinunn Root Criterion) were used to test the hypothesis
of no overall locality ellect, that is, no significant nujiphcjlogical
dilterence among samples, following a mnltivariate analysis of
variance ot all samples. All four criteria gave highly significant
/■-values (/^<().0()01 ), indicating that significant morpholc^gical
difterenc es exist among samples.

Tor the MANOV’A of all samples, \'ectors I to IV were signifi¬
cant and accounted for 65.21, 10.28, 5.92, and 4.69 per cent (tc3tal
86.10 per cent) of the variation, respectively. In the D. compactus
MA\()\'A, V ectors I and II accounted for 70.76 per cent of the
variation (53.48 and 17.28 per cent, respectively). Vectors I, II, and
III of the D. ordii MANOV'A explained 63.23 per cent of the vari¬
ation (26.93, 22.51, and 13.79 per cent, respectively).

Canonical analysis using all samples (fig. 7) depicts two
groups (A and B) which are comparable to the two clusters shown
in Fig. 6. Group A includes samples of D. compactus. As in pre¬
vious analyses, samples from the zone of sympatry separate readily
from one another. The major separation between the two groups
occurs along \'ector I. Several cranial features (GSL, GSW, LMB,
and WMB) exert a high influence on this vector as does the exter¬
nal character TAL (Table 1). Except for GSL, in which there is
general overlap in both species, these characters are consistently
larger in ordii than in compactus (Tig. 3). Samples of ordu are
tlistinguished from one another along the second vector; GSL and
WMB are again important as is SD.

Samples 1-5 separate from other samples of ordii, as they did in
the cluster analysis. Sample 2 (Oklahoma), because of its less
inflated mastoid bulla and wider supraot cipital and interparietal
bones, is somewhat intermediate between samples of ordn from
northern Texas (1, 3-5) and those of compactus. However, this
intermediacy is not thought to represent phenotypit affinity
toward compactus.

Canonical analysis for samples of D. compactus (Tig. 8) delin¬
eates two groups, island aiul mainland, which are identical to the
two grc^ups in the cluster analysis. Sample 58 again shows affinity
with the island form, although it is somewhat intermc'diate
between the two groups. The separation along \’ector I is most
heavily influenced by the characters lAL, HkL, C»SL, LMB,
VV’iMB, and SD. Little differentiation occurs along \'ector II.

Canonical analysis of D. ordii samples (excluding 1-5) reveals a
segregation of samples into three geographic groups (kig. 9). The

12

OCCASIONAL PAPERS MUSEUM TEXAS TECH UNIVERSITY

VECTOR I 65.21%

P'lG. 7. — Projections of the first two canonical vectors for all samples of Dipod-
omys compactus (A) and D. ordii (B). For each locality, the cross is positioned
near the mean value for each sample in the character space; the ellipse surround¬
ing each cross represents one standard deviation around the mean. See Figs. 1 and
2 for location of samples.

western group (A) is composed of specimens from Trans-Pecos,
Texas, (samples 6, 7) and northern Chihuahua, Mexico (9, 11, 12).
Sample 26 (Aguascalientes, Mexico), which is not geographically
proximal, is associated with this group, but samples 8 (Big Bend
Basin, Texas) and 10 (Samalayuca, Chihuahua, Mexico), which
are geographically close, are not. The second, or Mesa del Norte
group (B), consists of samples from northern Mexico (except 9, 11,
12), sample 8, and all samples of south Texas ordii. The third
group (C) is comprised of samples from the Mesa Central of Mex¬
ico (21, 23, 24, 25, 27), and this group is identical to the Mesa
Central division of the cluster analysis. The western group segre¬
gates from the others along Vector I, with the characters TAL,
GSL, NT, LMB, and WMB having the highest per cent influence.
The Mesa del Norte and Mesa Central groups segregate along
Vector II, with TL, TAL, HFL, GSL, GSW, and NL exerting the
highest influence.

BArMC;ARnNI-:R and SCHMIDI.V — KANC.AROO RA IS

I ABi.F. 1. — Rigeni’alues for the first two cntionual variates and the percentage
influence of each character in distinguishin^e; samples of Dipoclomys coiiipac iiis
and D. orclii from Texas and Mexico. The relative importance of each original var¬
iable to a particular canonical variate is computed by multiplying the eigenvalue
by the median value of the dependent variable, summing all values for a particular
vector, and then computing the per cent of relative importance of each variable per

vector.

Cluir.u ter

Median

('.anon leal

\’aiiaie I

Canonical

\'ariate II

KiKenvalne

Per t ent

infhieiue

Eigenvalue

Per r ent

influeiue

Tl.

238.18

-0.0010111

6.84

0.0004682

2.36

lAl.

131.67

0.0035409

13.23

-0.0016785

4.65

HFL

37.21

-0.0013423

1.42

0.0081740

6.40

GSL

37.36

-0.0146814

15.56

0.0291253

22.90

MW

20.19

0.0012186

0.71

-0.0002819

0.13

LIW

12.87

0.0099736

3.64

-0.0219979

5.96

GSW

23.54

0.0139852

9.34

-0.0019678

0.97

RW

3.60

-0.0485440

4.97

0.0288871

2.19

NL

13.60

-0.0181189

6.99

-0.0298818

8.54

LM I R

4.87

-0.0410823

5.68

0.0623567

6.40

LSW

2.28

-0.0369883

2.38

0.0207834

0.99

I. .MB

15.36

0.0310106

13.52

-0.0240858

7.79

W.MB

10.18

0.0398195

11. .50

0.0692352

14.84

.SD

13.31

0.0111129

4.22

0.0.567317

15.89

Di.scriminant junction analysis. — The histogram for the dis¬
criminant function scores of the compactus-ordii comparison
clearly shows two distinct groupings (A and B) with no interme¬
diate specimens (Fig. 10). Fhe Mahalanobis value for this
comparison (1)^=1 07.9; Fis 167=298.8; P<0.001) is well above that
reported by Wilkins and Schmidly (1979) for comparisons between
three species of pocket mice from west Fexas. Fhis indicaies that
cornpactus and ordii may he distinguished without ambiguity
from one another by using morphological features. Furthermore,
there is no indication of hyhridi/ation or morphological interme¬
diacy between cornpactus and ordii where their ranges overlap.

Fhe frecjuency liistogram of the /.-stores for the comparison
between island and mainland populations of /). cornpactus shows
a general segregation into two groups (T and D), although there
is some overlap between them. Fhere are four instaiues of a spec¬
imen being assigned to a group other than the one to which it
belongs geographically. Once again specimens from locality 58
show more affinity to tlie island than to the mainland samples.
Two specimens of 58 combine with the island group, one is
intermediate between island and mainland groups, and one sepa-

14

OCCASIONAL PAPERS MUSEUM TEXAS TECH UNIVERSITY

rates with the mainland samples. One other mainland specimen
(from locality 53) combines with the island group, whereas one
island specimen (from locality 59) combines with the mainland
rats. The value for this comparison is 10.3 (F15 i3g=23.8;
P<0.001).

Taxonomic Conclusions

Utilizing univariate and multivariate statistical techniques,
samples of Dipodomys compactus and D. ordii from South Texas
are distinct from one another without evidence of hybridization.
This agrees with results of studies by Johnson and Selander
(1971), Brownlee (1973), Best and Schnell (1974), Stock (1974), and
Schmidly and Hendricks (1976). In South Texas, the two species
appear to be confined to the Rio Grande Plain (Fig. 2). D. com¬
pactus occurs on the eastern two-thirds of the mainland, whereas
D. ordii inhabits the western two-thirds. Their ranges narrowly
overlap throughout central south Texas, and three sites of sympa-
try have been identified in Jim Hogg and Zapata counties.

Two subspecies are recognized within D. compactus (Fig. 11).
D. c. compactus, comprising populations formerly referred to the
subspecies compactus, largus, and parvabullatus , occurs on Mus¬
tang and Padre Islands of Texas and the barrier islands of
Tamaulipas, Mexico. D. c. sennetti inhabits the eastern two-thirds
of the South Texas mainland. Sample 58 exhibits marked affinity
for the island subspecies. All other mainland samples exhibit
more affinity for one another than for island samples.

D. ordii from the southern portion of the range in Texas and
Mexico are aligned herein into eight subspecies (Fig. 12; exclud¬
ing D. o. pullus which was not examined). The subspecies extrac-
tus, rnedius, oklahomae, ordii, and richardsoni were studied only
for comparison with the more southern races. Except for extrac-
tus, no evidence to contradict the current taxonomic arrangement
of these subspecies was found. D. o. extractus exhibits intermedi¬
acy between D. o. ordii and D. o. obscurus, which raises questions

Eig. 8.— Projec tions of the first two canonical vectors for samples of Dipodomys
compactus. CToup A represents samples of island compactus and group B samples
of mainland compactus. .See Eig. 7 for an explanation of symbols.

Eig. 9. — Projections of the first two canonical vectors for samples of Dipodomys
ordii. Ehese groups represent samples from; A, west Texas and north Chihuahua,
Mexico: B, Big Bend Basin and Rio Grande Plain of Texas and Mesa del Norte,
Mexico; and C, Me.sa Central, Mexico. See Fig. 7 for an explanation of symbols.

VECTOR II 17.28%

BArM(;ARnNKR AND SCHMIDl.V— KAN(,AR(K) RAIS IT)

o

u

UJ

>

— H

-oe

VECTOR I 26.93 %

•05

16

OCCASIONAL PAPERS MUSEUM TEXAS TECH UNIVERSITY

ADNanoaiid

DISCRIMINANT SCORE

P'lc 10. — Frequency hi.stograms ol discriminant function analysis comparisons of Dipodomys (above: D. compactus, group A, versus D.
ordii, group B; below: island compactus, group C, versus mainland compactus, group D).

BAUMC.ARDNKR AND SCHMIDI A— KANCiAROO RAIS

17

Fig. 11. — Geographic distributions of the subspecies of Dipodomys compncius:
1, D. c. cornpactus', and 2, D. c. sennet ti.

as to its status, although no change is recommended liere. For a
further discussion, see Anderson (1972) and Baumgardner (1979).

D. ordii from south Fexas was assigned to the subspecies dxir-
ranti by Schmidly and Hendricks (1976). However, in both uni¬
variate and multivariate analyses, south Fexas ordii consistently
group with samples from northern Mexico and the Big Bent!
region of Texas (Figs. 6B, 9). Fhis combined group, which is
referred herein to I), o. ohscurus, includes the previously retog-
nized subspecies attenuatus, idounis, and ohscurus as well as
northern samples of durrauti. Its range includes the Mesa del
Norte of Mexico and the adjacent regions of the Big Bend Basin
and Rio Grande Plain of Texas.

The remaining subspecies of I), ordii occur on the Mesa Gentral
of Mexico. D. o. durranti occupies the extreme northern portion
of this region in Nuevo Leon, Coahuila, San Luis Potosi, Famau-
lipas, and Zacatecus. D. o. palrneri inhabits the remainder of this
region south to Hidalgo. As noted by Schmidly and Hendricks

18

OCCASIONAL PAPERS MUSEUM TEXAS LECH UNIVERSITY

Fig. 12.— Geographic clistribution.s of the subspecies of Dipodomys ordu\ 1, D.
o. durranti; 2, D. o. extractus; 3, D. o. tyiedtus', i, D. o. obscurus; 5, D. o. oklaho-
mae; 6, D. o. ordii; 7, D. o. palmeri; and 8, I), o. richardso7ti.

(1976), the distinctiveness of these two subspecies is questionable.
They group closely together in the multivariate analyses (Fig. 6B,
9), although there are significant differences between them in
three of the five characters examined in the univariate analysis
(Fig. 5).

BAUM(;ARI)\KR and SCHMIDIA — KANC.AROO RAIS

III

SVSTKM A nC AcCX)L1 NTS

Only specimens used in statistical analyses are listed in the
accounts beyond. .Additional specimens, (onsistint^ t)f non-adults,
skin or skull only specimens, or unmeasured specimens, are listed
in Baumgardner (1979). The number in parenthesis preceding a
locality is its sample mmdx'r. The number of specimens from a
locality and the abbreviation of the institution of clej)osition fol¬
lows each locality.

Dipodomys compactus True

1889. Dipudoinys compactus ri iu% Proc. U.S. Nat. Mus., 1 1:160, January 5.

1891. Dipodops sennetti J. .Alien, Bull, .\niei. Mus. Nat. Hist., 3:226, .April 29,
type from near Brownsville, Claineron Co., Texas (part, specimens from the
eastern two-thirds of the .South Texas mainland).

1912. Dipodomys ordii, Davis, J. Mamm., 23:332, .August 14.

1976. Dipodomys compactus, Schmidly and Hendricks, Bull. .Southern C^alifornia
.Acad. .Sc i., 75:235, November.

Holotype. — None designated, but Poole and Schantz (1942:406)
assumed it to be a female, IkSNM 19665 35227; Padre Island,
Ciameron Co., Texas; 3 Ajjril 1888; obtained from C. K. Worthen.

Distribution. — The barrier islands c^f I'amaulipas, Mexico,
Mustang and Padre islands of South Texas, and the adjacent two-
thirds of the mainland east from Za[)ata County and south from
Bexar and Gonzales counties.

Comparisons. — See account of D. ordii.

Dipodomys compactus compactus True

1889. Dipodomys compac tus True, Pick. l'..S. Nat. .Mus., 1 1:160, January 5.

1942. Dipodomys ordii compactus, Davis, J. .Mamm., 23:332, .August 14.

1951. Dipodomys ordii parvahullatus Hall, IJniv. Kansas Puhl., Mus. Nat. Hist.,
5:38, October 1; tyjK' from 88 mi. S, 10 mi. \V Matamoros, Tamaulipas,
.Mc'xico.

1951. Dipodomys ordii lart^us Hall, Ihiiv. Kansas Puhl., Mus. Nat. Hist., 5:40,
October 1; type from .Mustang Island. 14 mi. .S\V Pori .Aiansas, .Aransas
Cio., Texas.

1976. Dipodomys compac tus compat tus, .Schmidly and Hendricks, Bull. Soulhein
California .Acad. .Sci., 75:235. .Novc-mhet.

1976. Dipodomys compactus lari^us, .Schmidly and Hcmdiicks, Bull. .Soulhein
Cialifornia .Acad. Sc i., 75:235. .November.

1976. Dipodomys compactus parcabullatus, .Schmidly and Hendiicks, Bull.
.Southern California .Acad. Sci., 75:235, November.

Holotype. — See account of D. compactus.

Distribution. — Mustang and Padre Islands of .Soutli lexas and
the barrier islands of Tamaulipas, Mexico.

20

OCCASIONAL PAPERS MUSEUM TEXAS TECH UNIVERSITY

Comparisons. — From D. c. sennetti, D. c. compactus differs in
having less inflated mastoid bullae, a slightly narrower skull with
wider supraoccipital and interparietal bones. Also, compactus
exhibits two dorsal pelage color phases (red and gray), whereas
sennetti has only the red phase.

Specimeyis analyzed (71). — Mexico: Tamaulipas; (22) 88-90 ini. S, 10 mi. W
Matamoros, 5 (KU); (22) 90 mi. S, 10 mi. W Matamoros, 1 (USNM). Texas:
Cameron Co.: (59) 2 mi. E, 6.5 mi. N Port Isabel, 2 (UIMNH); (59) Padre Island, 2
mi. E, 6 mi. N Port Isabel, 2 (UIMNH); (59) Padre Island, 3 mi. E, 6 mi. N Port
Isabel, 12 (TCWC); (59) 2 mi. E, 5 mi. N Port Isabel, 8 (UIMNH). Kleberg Co.:
(38) Padre Island, 10 (1 MWU, 9 USNM). Nueces Co.: (33) Mustang Island, SW
Port Aransas, 3 (TCWC); (33) Mustang Island, 1 mi. S Port Aransas, 3 (TNHC);
(33) 14 mi. SW Port Aransas, 5 (KU); (33) 15 mi. SW Port Aransas, Mustang
Island, 2 (TCWC); (33) 19 mi. S Port Aransas, Mustang Island, 16 (7 MV^Z, 9
TCWC); (36) 3.6 mi. S Bob Hall Pier on Padre Island, 1 (TAIU); (36) 23 mi. S
Port Aransas, 1 (TCWC).

Dipodomys compactus sennetti (J. A. Allen)

1891. Dipodops sennetti J. A. Allen, Bull. Amer. Mus. Nat. Hist., 3:226, April 29
(part, specimens from eastern two-thirds of the South Texas mainland).

1942. Dipodomys ordii sennetti, Davis, J. Mamm., 23:332, August 14 (part, spec¬
imens from the eastern two-thirds of the South Texas mainland).

1976. Dipodomys compactus sennetti, Schmidly and Hendricks, Bull. Southern
California Acad. Sci., 75:235, November.

Holotype. — Male, AMNH 3478/2733; near Brownsville,
Cameron Co., Texas; 9 March 1888; obtained by J. M. Priour;
type locality reported by Bailey (1905:45) as Santa Rosa, 85 mi.
SW Corpus Christi.

Distribution. — The eastern two-thirds of the South Texas main¬
land, east of Zapata County and south of Bexar and Gonzales
counties.

Comparisons. — See account of D. c. compactus.

Remarks. — Topotypes of this subspecies are the most atypical of
the mainland specimens and show affinity to the island
subspecies.

Specimens analyzed (83).— Texas: Atascosa Co.: (28) 7 mi. E Lytic, 1 (TNHC).
Brooks Co.: (53) Encino Division, King Ranch, near Encino, 2 (TCWC).
Cameron Co.: (58) near Santa Rosa, 4 (USNM). Hidalgo Co.: (53) 13.2 mi. S
Encino (Brooks Co.), Hwy. 281, 1 ( LCWC); (53) 4.4 mi. N Linn, Hwy. 281, 1
(ICWC). Jim Hogg Co.: (43) 1 mi. E, 1.2 mi. S Hebbronville, Hwy. 1017, 1
(TCWC); (43) 2 mi. S, 3.7 mi. W Hebbronville, 4 (TCWC); (41) 13.4 mi. SSE
Mirando City (Webb Co.), Hwy. 649, 21 (TCWC); (43) 7.2 mi. S Hebbronville,
Hwy. 1017, 1 (IC;WC); (44) 18.5 mi. SSE Mirando City (Webb Co.), Hwy. 649, 1
(TCWC); (43) 1 mi. E, 12.5 mi. S Hebbronville, Hwy. 1017, 1 (TCWC); (51) 20 mi.
S Hebbronville, 14 (TNHC); (51) 23.6 mi. S Hebbronville, Hwy. 1017, 3 (TCWC);
(51) 28.7 mi. S Hebbronville, Hwy. 1017, 2 (TCWC). Kenedy Co.; (46) 2.2 mi. S

bai'M(;ari)ner and sc:hmidi.v— kanc.aroo rais

21

Miflin, V.S. 11, 1 (TAIU); (Hi) 12 mi. S Sarita, 1 ( ICVVC); (51) 6 im. S Nonas,
Hwy 77, 1 ( 1 CAVC); (5-1) 1 mi. F. Rudolt, Noiias Division, King Ranch, 2
( rCVVCl); (5-1) 7 mi. F Rudolf, Norias Division, King Ranch, 1 ( rdWCl); (54) 8.6
mi. S Nonas, Mwy. 77, 2 (TCWC). Ki.kberg Co.: (46) 2 mi. S Riviera, 7 (TCVVC:).
Starr Co.: (51) 11.7 mi. F, 27.5 mi. N Rio Chande City, llvvy. 1017, 2 (I'CVVC).
Webb Co.: (37) 4 mi. WNW Bnmi, Hwy. 359, 1 ( rCiWC); (40) 9.1 mi. S Mirando
(.ity, Hwy. 649, 1 ( I'CVVC). Willacy Co.; (54) 6.2 mi. N Rayniondville, Hwy. 77, 1
( rCWCi); (56) Red Fish Bay, 28 mi. F Raymondville, 2 (TCWC); (56) Sauz Ranch,
2 (CSNM). Zapata C'.o.: (49) 2 mi.NF Bustamante, Hwy. 16, 2 (TCWC).

Dipodomys ordii VVoodhouse

1853. D[ tpodomys] ordn Woodhouse, Pick. Acad. Nat. Sci. Philadelphia, 6:224.

Holotype. — None designated. Species characterized from speci¬
mens obtained at El Paso, Texas, by Dr. Woodhouse.

Distribution. — From southern Canada to the southern edge of
the Mexican Central Plateau and from the eastern boundry of the
Rocky Mountains to the eastern limits of the Great Plains of the
Tnited States (Hall and Kelson, 1959).

Comparison. — In the southern range of D. ordii, compactus is
the only other five-toed kangaroo rat of comparable size. Exter¬
nally ordii differs from compactus in having a longer, bushier,
and slightly more crested tail. Also the ventral pencil is darker,
less broken, and extends to the tip of the tail. The pelage of ordn
is longer and silkier. In areas of proximity, the pelage of ordn
tends toward brownish hues whereas that of compactus has an
orange cast.

CTanially the two differ in the inflation of the mastoid bullae.
Ehat of ordn exhibits greater inflation, giving the skull a more
triangular appearance. This inflation causes the intermediate
supraoccipital and interparietal elements to be narrower. Also, the
interparietal of ordn comes to a finer point posteriorly and is
more triangular in shape than that of compactus, which is rec-
tangtdar to roundish in shape.

Dipodomys ordii durranti Setzer

1949. Dipodomy.s ordii fu.scus Set/cr, Tiiiv. Kansas Publ., Mus. Nat. Hist., 1:555,
December 27.

1952. Dipodotny.% ordii durranti .Sel/er, J. Wasbiiigtoii Acad. .Sci., 42:391,
December 17.

Holotype.— \du\t, male, TSNM 93886; Jaumave, 'Eamaulipas,
Mexico; 3 June 1898; obtained by E. W. Nelson and E. A.
Goldman.

22

OC;CASIONAL PAPERS Ml^SEUM TEXAS TECH UNIVERSEI Y

Distribution. — The northern half of the Mesa Central of Mexico
in southern Nuevo Leon and the adjacent regions of the states of
Coahuila, San Luis Potosi, Tamaulipas, and Zacatecus.

Comparisons. — See Setzer (1949).

Remarks. — Although previous authors have included specimens
from northern Tamaulipas and Coahuila, Mexico (Setzer, 1949)
and southern Texas (Schmidly and Hendricks, 1976) in this sub¬
species, our analyses indicate these samples should be referred to
D. o. obscurus.

Specimens analyzed (83). — Mexico; Coahuila: (21) 7 mi. S, 4 mi. E Bella Ihiion,
7200 ft., 28 (KLI); (21) 12 mi. W San Antonio de las Alazanas, 6500 ft., 2 (KlI); (23)
8 mi. N La Ventura, 5500 ft., 3 (REl); (23) San Juan Neponuceno, 5 mi. N La
Ventura, 4 (MVZ); (23) La Ventura, 6 (USNM). Nuevo Leon: (24) Dr. Arroyo, 1
(LISNM). San Luis Potosi: (24) 7.6 mi. S Matehuala, 2 (MVZ). Tamaulipas: (25)
Miquihuana, 10 (6 IISNM, 4 KU); (24) Nicolas, 56 km. NVV Tula, 5500 ft., 6 (KU);
(25) Juamave, 8 (I'SNM); (25) 8 mi. N Tula, 4500 ft., 2 (KU); (23) 3 mi. N Lulu, 3
(MVZ); (23) Lulu, 8 (MVZ).

Dipodomys ordii obscurus (J. A. Allen)

1891. Dipodops senyielti J. A. Allen, Bull. Amer. Mus. Nat. Hist., 3:226, April 29
(part, specimens from western two-thirds of the South Texas mainland).

1903. Perodipus obscurus J. A. Allen, Bull. Amer. Mus. Nat. Hist., 19:603,
November 12.

1921. Dipodomys ordii obscurus, Grinnell, J. Mamm., 2:96, May 2.

1939. Dipodomys ordii allenualus Bryant, Occas. Papers, Mus. Zook, Louisiana
State Ihiiv., 5:65, November 10, type from Mouth of Santa Helena Canyon,
2146 ft.. Big Bend of Rio CTande, Brewster Co., Texas.

1942. Dipodomys ordii sennetti, Davis, J. Mamm., 23:332, August 14 (part, spec¬
imens from the western two-thirds of the South Texas mainland).

1949. Dipodomys ordii idoneus Setzer, Univ. Kansas Publ., Mus. Nat. Hist.,
1:546, December 27, type from San Juan, 12 mi. VV Lerdo, 3800 ft.,
Durango, Mexico.

1976. Dipodomys ordii durranti, Schmidly and Hendricks, Bull. Southern Cali¬
fornia Acad. Sci., 75:2'55, November (part, specimens from northern
Eamaulipas and Coahuila, Mexico, and South Texas).

Holotype. — Adult, male AMNH 20957; Rio Sestin, northwestern
Durango, Mexico; 13 April 1903; obtained by J. H. Batty.

Distribution. — Northern portion of the Mexican Plateau above
southern Coahuila, known as the Mesa del Norte, and the adjoin¬
ing regions of the Big Bend Basin and Rio Grande Plain of
Texas.

Comparisons. — From D. o. durranti, D. o. obscurus differs in
having a slightly wider maxillary width and less inflated mastoid
bullae.

BArM(;ARnNKR AND SCiUMIDlA— kAN(;AR()() RAIS

2:^

From /). o. extractus, obscurus (litters in liaving a stiglitiy
smaller skull length, maxillary width, and less inllated mastoid
bullae.

From 1). o. ordii, obscurus differs in having a shorter skull,
narrower and shorter mastoid bullae, lesser skull depth, and a
slightly narrower maxillary width.

From /). o. palrneri, obscurus differs in having a greater skull
length, maxillary width, and skull depth as well as a slightly
longer and narrower bullae.

Specimens analyzed (203). — Mkxico: CIhihuahua: (14) Las Areiiosos, 4050 fl., 6
(KL); (14) Sierra .Alinagre, 5300 ft., 12 mi. S Jaco, 6 (Klh: (17) 15 mi. ESE
Boquilla, 4700 ft.. 2 (RL): (17) 19 mi. N. 7 mi. E Parral, 1 (KlI); (17) 2 mi. E La
Parrena, 5000 ft., 1 (KL); (17) 5 kin. S Jiminez, 2 (Kl’); (17) 5 mi. E Parral, 5700
ft.. 7 (KL). CoAttuii.A: (13) 11 mi. W Ilcda. .San Miguel, 2200 ft., 1 (Kl'); (18) 3 mi.
N'E Sierra .Mojatla, 4100 ft., 1 (KL): (16) 1 mi. S Hermanas, 1 (KL): (18) 4 mi. N
.\catita, 3600 ft., 2 (KL); (18) 1 mi. S\V San Pedro de las Colonias, 3700 ft., 3 (KL);
(18) 8 mi. SE .San Pedro de las Colonias, 3700 ft., 1 (KL); (18) 10 mi. E Torreon,
3600 ft., 6 (KL): (18) 1 mi. N San Lorenzo, 4200 ft., 3 (Kll); (20) N foot Sierra
(iitadalupe, 6200 ft., 9 mi. S, 5 mi. \V Cieneral Opeda, 7 (KL). Durango: (19) Rio
de Bocas, 7 (A.MNH); (18) 1 mi. \VS\V Mapirni, 3800 ft., 3 (KL); (19) Rosario. 4
(A.MNH); (18) 5 mi. SE Lerdo, 3800 ft.. 5 (KL); (19) 6 mi. N\V Rodeo, 4200 ft.. 1
(KL). rAMAUi.iPAS: (15) Nuevo Laredo, 3 (IkSNM). Texas; Brewster CM.: (8)
Cooper’s Well, 47 mi. S Marathon, 2450 ft., 3 (M\’Z): (8) Big Bend National Park
(BBNP), 10 mi. NE Panther Junction, 2820 ft.. 4 (SWESl’MC); (8) BBNP, Lpper
Eornillo Clreek Bridge. 1 (SW TSl’MC:); (8) Lpper 'Eornillo Creek Bridge, BBNP, 8
mi. NNE Panther Junction, 14 ( ECAV’C;); (8) Lower 'Eornillo Creek Bridge, BBNP,
15 mi. SE Panther Junction, 1 ( ECVVC); (8) Mouth Santa Helena (Canyon, 2146 ft..
Rio C.rande, 2 (1 ECWCi, 1 M\'Z). Dim.mit Co.; (30) 2 mi. NE Carrizo Spring
along Nueces River, 1 ( ECAV’C); (30) 10 mi. SW Carrizo Springs, 1 ( ECAVC); (30) 2
mi. SW .\sherton, Hwy. 1916, 2 (TCWC). Hidaigo Co.: (57) 17 mi. NW Edinhurg,

I ( ENHC). Jim Hogg Co.: (42) 13.4 mi. .SSE Mirando City (Wehh Co.). Hwv. 649,

II (TCiWC;); (42) 14 mi. .SSE .Mirando City (Wehh Co.). Hwy. 619, 1 (TCAVC); (42)
14.3 mi. SSE .Mirando City (Webb Co.), Hwy. 649, 1 ( ECWC); (42) 14.7 mi. SSE
.Mirando Caty (Webb Co.), Hwy. 649, 3 ( ECAVCi); (42) 16.1 mi. .SSE Mirando City
(Webb C:o.). Hwy. 649, 2 (LCAVC): (45) 18.1 mi. SSE Mirando C;iiy (Webb Co.).
Hwy. 649, 1 ( ECWC); (45) 18.7 mi. SSE .Mirando City (Webb Co.). Hwv. 619. 2
(TCWC); (45) 19.4 mi. SSE .Mirando City (Webb Co.). Hwy. 649, 4 ( KAVC); (45)
20 mi. .SSE .Mirando City (Webb (io.). Hwy. 649. 2 (I(iWC): (45) 20.3 mi. SSk
Mirando City (Webb Co.), Hwy. 649, 3 ( ECiWC); (47) 22.5 mi. SSE Miiando C.iiy
(Webb C;c).). Hwy. 649, 1 (TCWC;): (47) 23.7 mi. SSE .Mirando City (Webb (>>.).
Hwy. 649, 1 (ECWC); (47) 25.6 mi. SSE Mirando City (Webb Cet.). Hwy. 619. 2
(TC:WC:); (47) 14 mi. N. 3 mi. W (dieira. Hwv. 649, I ( ECWC); (47) 26 mi. SW
Hebbronville. Hwy. 16. 4 (ECWC); (47) 23 mi. S. He-bbietnville. Hwv. 16. 1
( ECAV’C:); (47) 18 mi. SW Hebbremville. Hwy. 16. 2 ( ECAVC); (47) 22 mi. SW Heb¬
bronville. Hwy. 16, 1 (ECWC:); (47) 20.5 mi. SW Hebbreinville. Hwy. 16. 3
( ECWC): (47) 13.7 mi. N (dterra, Hwy. 649, 2 ( E(AVC): (.52) 0.9 mi. N (.ue'iia,
Hwy. 649, 1 ( ECTVC:); (52) 2.8 mi.S Cdterra, Hwy. 649. 1 (T(AVC). Mc;Mt i.t ES Co.:
(29) 15 mi. NE Eilden. 1 ( ECWC). Webb CM.: (32) 40 mi. SW Catarina, em Rio

24

OCCASIONAL PAPERS MUSEUM TEXAS TECH UNIVERSITY

Grande, 2 (TNHC); (35) 21 mi. NE Laredo city limits, 1 (USFWS); (34) 21.1 mi. N
Brum, Hwy. 2050, 2 (TCWC); (35) 13 mi. NE Laredo city limits, 2 (USFWS); (35)
12 mi. NE Laredo city limits, 1 (USFWS); (34) 15 mi. N Aguilares, Hwy. 2895, 1
(TCWC); (35) 11 mi. NE Laredo city limits, 2 (USFWS); (35) 8 mi. NNE Laredo
city limits, 1 (USFWS); (35) 10 mi. NE Laredo city limits, 1 (USFWS); (35) 6.5 mi.
NNE Laredo city limits, 3 (USFWS); (35) 5 mi. NE Laredo city limits, 1 (LISFWS);
(35) 4 mi. ENE Laredo city limits, 1 (USFWS); (35) 5 mi. E Laredo city limits, 1
(IISFWS); (35) 8 mi. E Laredo city limits, 1 (USFWS); (35) 10 mi. E Laredo city
limits, 1 (LISFWS); (39) 2 mi. N Aguilares, Hwy. 2895, 1 (TCWC); (39) 4.5 mi. SSE
Mirando City, Hwy. 649, 2 (TCWC); (39) 5.6 mi. SSE Mirando City, Hwy. 649, 1
(TCWC); (39) 6.6 mi. SSE Mirando City, Hwy. 649, 2 (TCWC). Willacy Co.: (55)
10 mi. NW Raymondville, 5 (TNHC). Zapata Co.: (48) 10.9 mi. NE Bustamante,
Hwy. 16, 2 (TCWC); (48) 9.8 mi. NE Bustamante, Hwy. 16, 2 (TCWC); (48) 7.6 mi.
NE Bustamante, Hwy. 16, 1 (TCWC); (48) 2 mi. NE Bustamante, Hwy. 16, 3
(TCWC); (50) 5 mi. N Zapata, 1 (TNHC); (50) 3 mi. SW Bustamante, Hwy. 16, 3
(TCWC); (50) 4 mi. SW Bustamante, Hwy. 16, 1 (TCWC); (50) 3.5 mi. NE Zapata,
4 (TNHC).

Dipodomys ordii palmeri (J. A. Allen)

1881. Dipodops ordii palmeri J. A. Allen, Bull. Mus. Comp. Zool., 8:187, March.
1921. Dipodomys ordii palmeri, Grinnell, J. Mamm., 2:96, May 2.

Syntypes. — Two adult males, MCZ 5886 and 5887; San Luis
Potosi, Mexico; 1 May 1878 and 1 September 1878, respectively;
obtained by Dr. Edward Palmer.

Distribution. — Southern portion of the Central Plateau of Mex¬
ico from northern San Luis Potosi and Zacatecus south to
Hidalgo.

Comparisons. — See Setzer (1949).

Remarks. — D. o. palmeri differs from durranti in three of the
five univariate characters examined with Dice-Leraas diagrams;
however, these two subspecies group together consistently in the
multivariate analyses. For this reason, their subspecific distinct¬
ness is questionable, and additional study may show they are sim¬
ilar enough to be placed under the single subspecies palmeri,
which has priority over the name durranti.

Specimens analyzed (24). — Mexico: Aguascalientes: (26) 1 mi. N Chicaloie, 2
(MVZ). San Luis Potosi: (27) 2 mi. NW San Luis Potosi, 2 (MVZ); (27) Jesus
Maria, 18 (USNM). Zacatecus: (26) 4 km. E Morelos, 2 (MWLJ).

Other Subspecies

Subspecies accounts are not included for D. o. oklahomae, D. o.
richardsoni, D. o. medius, D. o. extractus, and D. o. ordii. Infor¬
mation presented in this study does not alter the accounts, de¬
scriptions, and distributions of these taxa as provided by Setzer
(1949). Specimens analyzed for these subspecies are as follows;

BAUMCiARDNKR AND SCHMIDl.V — KANC.AROO RAIS

2b

D. o. oklahomae ( 10).— Oklahoma: C;lkvkland Cio.: (2) 2.5 mi. S Norman, 10
(KU).

D. o. richardsoni ( 16).— I kxa.s: Floyd Oo.: (3) 21 mi. K Floydada, 1 ( FCWC).
Hartley Cio.; (1)1 mi. S\V Dalhart. 1 ( FCAVC); (1) 2 mi. S\V Dalhart, 1000 ft., 2
( rCVVO). Hemphill Oo.: Gtme Howe Wildlife Management Area: (1) Persimmon
Ciap, on creek, 10 mi. NE Clanadian, 2 (TCAVC); (1) 7.5 mi. NE Canadian, 1
(FCWC); (1) 7 mi. NE Canadian, 2 (FCAVC); (1) 5 mi. NE Canadian, 1 (TCWC);

(I) 6 mi. ENE Canadian, 2 (TCAV'C). Potler Co.: (1)2 mi. W Lake Meredith, 2700
ft., 2 (TCWC): (1) 18 mi. N Amarillo, 3.500 ft., 1 ( FCWC). Wheei.er Co.: (1) Wal¬
lace Ranch, SW Wheeler, 1 (TCWC).

D. o. rnednis (28). — Texas: Andrews Co.: (5) 15 mi. SW Andrews, 3000 ft., 1
(TCWC). Caines Co.: (4) Cedar Lake, 20 mi. ENE Seminole, 3 (FCWC). Ward
Co.; (5) 2 mi. NE Monahans, 3 (MWU); (5) 11 mi. W .Monahans, 2 (MWL^).
Winkler Co.: (5) 3.5 mi. S Kermit, 19 (TCWC).

D. o. ordii (85). — Mexico: Chihuahua: (9) 10 mi. SE Zaragosa, 3700 ft., 5 (KU);

(II) 1 mi. S Kilo, 4185 ft., 2 (KLI); (11) 8 mi. E V'illa Ahumada, 4000 ft., 2 (KL^);
(12) 11 mi. NNW San Buenaventura, 1 (KLI); (12) 1 mi. N Arados, 1540 m., 1 (KF');
(12) 2 mi. W Parrita, 2 (Kl^). Texas; Culberson Co.: (7) 16 (TCWC). El Paso Co.:
(6) 3 mi. NE El Paso city limits, 8 (MVZ); (6) 7.5 mi. E El Paso City Hall, 4000 ft.,
12 (KU); (6) 12 mi. E, 1 mi. S El Paso City Hall, 4000 ft., 4 (KU); (6) 18 mi. E, 3
mi. S El Paso City Hall, 4000 ft., 8 (KU); (6) 11 mi. SE El Paso City Hall, 2 (KU).
H udspeth Co.; (6) Fort Hancock, 6 (MWU). Jeff Davis Co.: (7) Limpia Creek, 16
mi. NE Fort Davis, 1 (KU). Presidio Co.: (7) 2 mi. S Paisano, 9 (TCWC); (7) 1 mi.
W Plata, 2 (MWLI); (7) Bandera Mesa, 2 (MWLJ); (7) 3 mi. E Presidio, 1 (MWL^).
Reeves Co.; (7) 20 mi. S Pecos, 1 (KU).

D. o. extractus (15). — Mexico; Chhiuahua: (10) 8 mi. NE Samalayuca, 4300 ft., 2
(KU); (10) 1 mi. E Samalayuca, 4500 ft., 13 (MV'Z).

Acknowledgments

Many people helped in the course of this research. Special
thanks are extended to Dr. Fred S. Hendricks of Texas A8cM Uni¬
versity for his assistance and evaluation of the manuscript. Por¬
tions of this paper were completed while the senior author was
employed by the Texas Natural History Collection, Texas Memo¬
rial Museum, Fhe University of Texas at Austin. Fhe assistance
of Dr. Robert F. Martin of that insiitution is gratefully
appreciated.

Institutions from which specimens were examined together with
institutional abbreviations icsed in the lists of specimens follow
(curators given in parenthesis): AMNH, American Museum of
Natural History (Syndey Ander.son): Klh Museum of Natural His¬
tory, The University of Kansas (Robert S. Hoffmann): MCZ,
Museum of Comparative Zoology, Harvard University (Barbara
Lawrence); MUZ, Museum of Vertebrate Zoology, Ihiiversity of
California (William Z. Lidicker, Jr.); MWU, Department of Biol¬
ogy, Midwestern University (Walter W. Dalquest); SVVISl’MC,

26

occ;asional papers museum texas tech universeey

Southwest Texas State University Mammal Collection, Southwest
Texas State University (John T. Baccus); TAIU, Department of
Biology, Texas A&I University (Allan H. Chaney); TCWC, Texas
Cooperative Wildlife Collection, Texas A&M University (David J.
Schmidly); TNHC, Texas Natural History Collection, Texas
Memorial Museum, University of Texas at Austin (Robert F. Mar¬
tin); UIMNH, Museum of Natural History, University of Illinois
(Donald F. Hoffmeister); USNM, National Museum of Natural
History, Bird and Mammal Laboratories, U.S. Bureau of Sports
Fisheries and Wildlife, Washington, D.C. (Clyde Jones and
Charles O. Handley); USFWS, U.S. Bureau of Sport Fisheries and
Wildlife, Denver Collection of the Bird and Mammal Laboratories
(Robert B. Finley). We are especially grateful to these individuals
who loaned us specimens or allowed us to study material under
their care.

This paper represents contribution No. TA 16005 of the Texas
Agricultural Experiment Station, Texas A&M University.

Literature Cited

Anderson, S. 1972. Mammals of Chihuahua; taxonomy and distribution. Bull.
Amer. Mus. Nat. Hist., 148:149-410.

Bailey, V. 1905. Biological survey of Texas. N. Amer. Fauna, 25:1-222.
Baumgardner, G. D. 1979. Systematics of the southern races of two species of
kangaroo rats (Dtpodomys compactus and D. ordi). Unpublished Mas¬
ters thesis, Texas A&M Univ., College Station, 96 pp.

Best, T. L., and G. D. Schnell. 1974. Bacular variation in kangaroo rats
(genus Dipodomys). Amer. Midland Nat., 91:257-270.

Brownlee, A. S. 1973. Differentiation of nine species of Dtpodomys (Rodentia:

Heteromyidae); a numerical taxonomy study based on morphology. Un¬
published Ph.D. dissertation, Univ. Mississippi, Ihiiversity, viii+81. pp.
Desha, P. G. 1967. Variation in a population of kangaroo rats, Dtpodomys
ordtt medtus (Rodentia: Heteromyidae) from the High Plains of Texas.
Southwestern Nat., 12:275-289.

Hall, E. R., and K. R, Kelson. 1959. The mammals of North America. Ron¬
ald Press Co., New York, 1 :xxx+546+79.

Hooper, E. T. 1952. A systematic review of the harvest mice (genus Retthrodortt-
omys) of Latin America. Misc. Publ. Mus. Zool., ETniv. Michigan,
77:1-255.

Honeycutt, R. L., and D. J. Schmidly. 1979. Chromosomal and morphological
variation in the Plains pocket gopher, Geomys bursartus, in Texas and
adjacent states. Occas. Papers Mus., Texas Tech Univ., 58:1-54.

J0HN.SON, VV. E., AND R. K. Selander. 1971. Protein variation and systematics in
kangaroo rats (genus Dtpodomys). Syst. Zool., 20:377-405.

Kennedy, M. L., and G. D. Schnell. 1978. Geographic variation and sexual
dimorphism in Ord s kangaroo rat, Dtpodomys ordtt. J. Mamm
59:45-59.

BAUMC.ARDNER AND SCHMIDI A — KAN(,AR()() RAIS

27

l.iDiCKKR, \V. Jr. 1960. An analysis ot intraspt'i ilic vaiialion in ilic kangaion
rat Diljodomys merriami. I'niv. California Publ. Zool., 67:125-218.

Long, C. 1968. An analysis of patterns ot variation in some representative
Mammalia, Part 1: A review of estimates of variability in selected mea¬
surements. Trans. Kansas Acad. Sci., 71:201-227.

Pooi.E, J., AND \ . S. ScMANTz. 1942. Catalog of the type specimens of mam¬
mals in the United States National Museum, including the biological
surveys collection. Bull. U.S. Nat. Mus., 178:1-705.

Rohi.k, F. J., and J. KtsttPAUGit. 1972. Numerical taxonomy system of multivar¬
iate statistical programs. The State Univ. New York at Stony Brook,
Stony Brook, New York. [“Manual” pi intout from program.]

ScH.MiDLY, D. J. 1971. Population variation in Dipodornys ordii from western
Texas. J. Mamm., 52:108-120.

Scn.MtDt.Y, D. J., AND F. S. Hendricks. 1976. Systematics of the southern races
of Ord’s kangaroo rat, Dipodornys ordii. Bull. Southern California
.\cacl. .Sci., 75:225-237.

Setzer, H. W. 1949. .Subspeciation in the kangarcro rat, Dipodornys ordii. Univ.
Kansas Publ., Mus. Nat. Hist., 1:473-573.

.Stock, .\. D. 1974. Chromosome evolution in the genus Dipodornys and its
phylogenetic implications. J. Mamm., 55:505-526.

Wilkins, K. T., and D. J. Sch.midly. 1979. Identificatic:)n and distribution of
three species of pocket mice (genus Perognathus) in Trans-Pecos Texas.
.Southwestern Nat., 24:17-32.

Yates, T. L., and D. J. Sch.midly. 1977. Systematics of Scalopus aquaticus
(Linnaeus) in Texas and adjacent states. Occas. Papers Mus. Texas Tech
Univ., 45:1-36.

.\ddresses of authors: Department of Wildlife and Fisheries Sciences, Texas A&M

I'nix'ersity, College Station, 778-13. Present address of Baumgardner; Ecological

Research Center, Department of Biology, Memphis State [Sin>ersity, Memphis,

Tenries.see 38132. Received 21 April, accepted 22 August 1980.

PUBLICATIONS OF THE MTSETM
I'EXAS TECH UNIVERSITY

Ehrec publications of The Museum of Texas Tech Ihii-
versity are issued under the auspices of the Dean of the
Ciraduate School and Director of Academic Publicaticms, and
in cooperation with the International Center for Arid and
Semi-Arid Land Studies. Short research studies are jitiblished
as Occasional Papers whereas longer contributiems appear as
Special Publications. Papers of practical application to col¬
lection management and museum operations are issued in
the Museology series. All are numljered separately and pub¬
lished on an irregtilar basis.

The preferred abbreviation for citing The Museum’s Occa¬
sional Papers is Occas. Papers Mus., Texas Tech Ihiiv.

Institutional subscripticjns are available through Texas
Lech Press, lexas Tech llniversity, Lubbock, Texas 79409.
Institutional libraries interested in exchanging publications
should address the Exchange Librarian at Texas Tech Ihii-
versity. Individuals can purchase separate numbers of the
Occasional Papers for $1.00 each from Texas Tech Press.
Remittance in IhS. currency check, money order, or hank
draft mtist Ix' enclosed with request (add $1.00 per title or
200 pages of publications requested for foreign postage; resi¬
dents of the state of Texas must pay a 5 jxa cent sales tax on
the total purchase price). Copies of the “Revised checklist of
North American mammals north of Mexico, 1979” (Jones et
al., 1979, Occas. Papers Mus., Texas Tech Univ., 62:1-17) are
available at 60 cents each in orders of 10 or more.

ISSN 0149-175X
Texas Tech Press
Ltthbock, Texas 79409

r^x

OCCASIONAL PAPERS
THE MUSEUM
TEXAS TECH UNIVERSITY

'•Uo. COMP. ZOOU.

library
NOV 1 fci 1981

' " A O

Nl'MBER 74

30 OC rOBER 1981

MICE OF THE GENUS PEROMYSCUS
IN GUADALUPE MOUNTAINS NATIONAL PARK, TEXAS

John E. Cornely, David J. Schmidly, Hugh H. Genoways,

AND Robert J. Baker

Mice of the genus Peromyscus are found in virtually every habi¬
tat type in Guadalupe Mountains National Park in West Texas.
Because of their abundance and wide distribution, they comprise
an important component of the park’s ecosystem. The first known
specimens of Peromyscus from the area now included in the park
were collected by Vernon Bailey in 1901 (Bailey, 1905). He col¬
lected specimens of Peromyscus boylii in Dog and McKittrick
canyons. Davis (1940) collected P.leucopus at Frijole in 1938 and
P. boy III in The Bowl in 1938 and 1939. Davis and Robertson
(1944) reported collecting P. pectoralis from along Bell Creek in
1938. A previously unreported specimen of P. difficilis collected in
1901 by V’ernon Bailey in McKittrick Canyon was recently
reported by Diersing and Iloffrneister (1974).

During a survey of mammals of Guadalupe Mountains
National Park conducted from June 1973 to August 1975, speci¬
mens of seven species of Peromyscus were obtained. In addition to
the four previously reported sjjecies, specimens of P. eremicus, P.
maniculatus, and P. truei were collected. We are not aware of any
other place where seven species of Peromyscus occur in such a
small area.

Genoways et al. (1979) presented general distributional, ecologi¬
cal, and faunal information for all mammals occurring in Guada¬
lupe Mountains National Park. In the present report the use of
morphometries and karyology as methods of identification of deer
mice in the park are discussed, and the distribution and ecology